Implications of Byproduct Chemistry in Nanoparticle Synthesis | The

11 Jun 2019 - Subscribed Access ..... These reactions are likely to occur in our MES–AuNP sols and therefore contribute to the byproduct formation. ...
0 downloads 0 Views 5MB Size
Article Cite This: J. Phys. Chem. C XXXX, XXX, XXX−XXX

pubs.acs.org/JPCC

Implications of Byproduct Chemistry in Nanoparticle Synthesis Frederick N. Stappen,† Kasper Enemark-Rasmussen,† Glen P. Junor,‡ Mads H. Clausen,†,§ Jingdong Zhang,† and Christian Engelbrekt*,†,∥ †

Department of Chemistry, Technical University of Denmark, Kemitorvet 207, 2800 Lyngby, Denmark Department of Chemistry and Biochemistry, University of California, San Diego, La Jolla, California 92093-0358, United States § Center for Nanomedicine and Theranostics, Kemitorvet 207, 2800 Lyngby, Denmark ∥ Department of Chemistry, University of California Irvine, Irvine, California 92697-2025, United States Downloaded via NOTTINGHAM TRENT UNIV on August 17, 2019 at 10:20:57 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.



S Supporting Information *

ABSTRACT: Byproducts in metal nanoparticle synthesis can interfere with nanomaterial formation and self-assembly, as well as the perceived nanomaterial properties. Such syntheses go through a complicated series of intermediates, making it difficult to predict byproduct chemistry and challenging to determine experimentally. By a combined experimental and theoretical approach, the formation of organic byproducts is mapped out for the synthesis of gold nanoparticles with Good’s buffer 2-(N-morpholino)ethanesulfonic acid. Comprehensive nuclear magnetic resonance studies supported by mass spectrometry, ultraviolet−visible spectroscopy, and density functional theory reveal a number of previously unidentified byproducts formed by oxidation, C− N bond cleavage, and C−C bond formation. A reaction mechanism involving up to four consecutive oxidations is proposed. Oligomeric products with electronic transitions in the visible range are suggested. This approach can be extended broadly and lead to a more informed synthesis design and material characterization.



INTRODUCTION

in nanomaterial synthesis, which may lead to incorrect conclusions based on the behavior of nanomaterial mixtures.11 Meanwhile, the drive for green syntheses of gold nanomaterials has led to the successful use of benign compounds as reducing and stabilizing agents12−16 and resulted in a plethora of materials with wide-ranging properties. Generally, green synthesis of AuNPs is done by mixing tetrachloroaurate [AuCl4]− with a reducing agent (e.g., NaBH417 or glucose2) and a stabilizing agent (e.g., starch2), or molecules acting as both, for example, 2-(N-morpholino)ethanesulfonic acid (MES),11 citrate,18 tryptophan,14 or aniline.19,20 Green AuNPs are generally evaluated via their optical properties and morphology through ultraviolet−visible (UV−vis) spectroscopy and transmission electron microscopy (TEM), respectively. The properties are typically attributed to the gold nanostructures and sometimes the stabilizing agent.13−15 Several reports on the synthesis of gold nanoclusters (AuNCs) using heteroatom-rich small organic reducing agents present fluorescence as the main support for AuNC formation,8−10 while these transitions could be assigned to organic byproducts.11 One appealing reagent in AuNP synthesis is MES, as it serves multiple purposes simultaneously: reduction, buffering, and capping. Furthermore, it is frequently used in biochemistry

Gold nanoparticles (AuNPs) are among the most studied nanomaterials and central to much research in nanotechnology and nanomedicine.1 Understanding the formation of both nanoparticles and organic byproducts is critical for the rational design of the nanomaterials, particularly for health care and food industries.2−4 As an example, the toxicity of nanomaterials has, in some cases, been linked to impurities deposited on the nanomaterial surface, which, in turn, influences the evaluation of the biocompatibility of nanomaterials and makes such studies very sensitive to purification methods.5 Furthermore, knowing the reaction mechanisms, and thus the organic products formed en route to the inorganic nanomaterials, will clarify the role of different components on behavior and properties of the systems. For instance, an organic byproduct was found to be responsible for the formation of superlattices during AuNP synthesis with t-butylamine,6 while macroscale AuNP-coated fibers were formed from trimethylamine oxidation byproducts,7 highlighting the potential of byproducts to dictate the nanomaterial structure. Thus, understanding byproduct formation allows for more rational design and manipulation of nanomaterial syntheses. Finally, molecularsized nanomaterials, such as nanoclusters, cannot be easily separated from organic byproducts and are thus often applied without rigorous purification.8−10 Since nanoparticle formation commonly occurs through nonselective redox chemistry, byproduct characterization is difficult and often overlooked © XXXX American Chemical Society

Received: April 5, 2019 Revised: June 7, 2019 Published: June 11, 2019 A

DOI: 10.1021/acs.jpcc.9b03193 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C as a nontoxic and optically inactive buffer and is generally considered inert. MES is used as a buffer in biological AuNPmediated H2O2 detection, where HAuCl4 is reduced by H2O2 and the formation of AuNPs is detected by UV−vis spectroscopy.21−24 Even though it is often recognized that MES may itself reduce HAuCl4, the influence of the byproducts of this side reaction on detection or the biological systems themselves have not been considered.21−24 One possible reason for the dearth of mechanistic studies is the complex progression of gold intermediates, making it difficult to predict the fate of the organic reducing agent, as well as its influence on the optical (and other) properties of the AuNP sols.2,25 Complicating the matter, all intermediates in gold nanostructure synthesis, including gold complexes,26 nanoparticles,27,28 and nanoclusters,29 are known catalysts for reactions, such as the oxidation of amines to imines,30 oxidations with dioxygen,27 and CO oxidation.31 Despite the challenge, organic byproducts formed during AuNP synthesis are sometimes discussed16 and can be supported by infrared (IR) absorption spectra, though the complexity of the resulting mixtures makes it difficult to reach firm conclusions.15,32 Newman and Blanchard studied AuNP formation with amine reducing agents by NMR spectroscopy and identified polyaniline formation after the reduction of HAuCl4 with aniline using 1H NMR.20 Similarly, dense organic layers were observed around AuNPs when HAuCl4 was reduced by tryptophan, and polytryptophan from oxidative polymerization was suggested based on 1H NMR and UV−vis spectroscopies.12,14 However, the sensitivity and resolution of standard 1H NMR spectroscopy can be a limiting factor for looking at byproduct mixtures at low concentrations. Regardless of the difficulties in characterizing organic byproducts in nanomaterial synthesis, such efforts are necessary because (1) byproducts may interfere with the perceived properties of the nanomaterials, (2) byproducts may associate strongly with the nanomaterial surfaces affecting their toxicity, and (3) byproducts can play an active role in the evolution of nanomaterial morphology. The current work presents a detailed characterization of the organic byproducts and chemistry involved in gold nanoparticle synthesis, as well as the effects of byproducts on the properties of nanoparticle sols. The MES-stabilized AuNPs studied here are conveniently removed by syringe filtration, confirmed by UV−vis spectroscopy and inductively coupled plasma optical emission spectroscopy (ICP-OES).11 We conduct a comprehensive NMR study supported by UV−vis spectroscopy and mass spectrometry (MS) to examine the byproducts from HAuCl4 reduction by MES. A range of MES derivatives are identified, and a reaction mechanism with up to four consecutive oxidation steps is proposed. The electronic absorption spectra of identified intermediates and oligomers are simulated by time-dependent density functional theory (TD-DFT) and molecular dynamics (MD) simulations, supporting the formation of conjugated dimers and radicals.

Figure 1. Synthesis, purification, and nanostructure of AuNPs synthesized at (a−c, f) 21 °C and (d, e, g) 95 °C. (a) Photographs at different reaction times during the first 6 min of AuNP synthesis at 21 °C. (b, d) Comparison of as-synthesized sol (left) and filtrate (right). (c, e) Syringe filtration showing the retention of AuNPs from the sol (in the syringe) by the filters. (f, g) TEM micrographs of AuNP product from synthesis at (f) 21 °C and (g) 95 °C.

mM MES), and high MES (2 mM HAuCl4, 75 mM MES) concentrations. After a certain reaction time, the AuNPs were removed by filtration through two sequential 0.2 μm polypropylene syringe filters (VWR international, Figure 1c,e) onto a Petri dish or watch glass, where the sample was kept at 80 °C until dry. The dried powder was redissolved in H2O or D2O for NMR or chromatographic analysis. The efficiency of AuNP removal by the filtration method was verified by UV−vis spectra of the filtrates showing no detectable AuNP absorbance and ICP-OES showing >99.995 and >99.974% removal of gold (corresponding to ∼0.08 and 0.5 μM, Figure S4) for syntheses carried out at 21 and 80 °C, respectively, in agreement with the previous results.11 NMR Spectroscopy. The presented NMR data were recorded on a Bruker AVANCEIII spectrometer operating at a 1 H frequency of 799.75 MHz (Oxford Instruments superconducting magnet refitted for Bruker Biospin electronics). The system was equipped with a 5 mm TCI CryoProbe (Bruker Biospin), and all experiments were performed at 298 K. Filtrates were redissolved in 600 μL D2O, unless otherwise noted. Ultraviolet−Visible (UV−Vis) Spectroscopy. UV−vis spectra were recorded on an Agilent 8453 (Santa Clara, CA) or PerkinElmer Lambda 950 (Waltham, MA) spectrophotometer in quartz cuvettes. In situ UV−vis experiments were performed by mixing HAuCl4 and MES in a glass vial at room temperature with stirring for 4 min before transferring the solution to a 1 mm quartz cuvette.



METHODS Synthesis of Gold Nanoparticle Sols. AuNPs were synthesized in a 100 mL round-bottom flask under magnetic stirring in an oil bath by adding a 20 mM HAuCl4 stock solution to MES in water at the target temperature (Figure 1a and Video S1).11,33 The concentrations of HAuCl4 and MES were varied. Three conditions were examined, i.e., standard (2 mM HAuCl4, 10 mM MES), high gold (10 mM HAuCl4, 10 B

DOI: 10.1021/acs.jpcc.9b03193 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

Figure 2. Observation of byproduct dynamics following AuNP formation at 21 °C. (a) Evolution of the extinction spectrum of the AuNP sol from 4 to 300 min. (b) Absorbance of AuNP-free filtrates prepared after 20−160 min of reaction. (c) Difference spectra generated by subtracting the initial spectrum, 4 min, in (a) from the following spectra, revealing a series of absorptions bands. (d) Time-dependent intensity of peaks labeled in (c). (e) Evolution of the absorption spectrum of the filtrate when AuNPs are physically removed by filtration after 3 min. (f) Time-dependent intensity of peaks labeled in (e).

Chromatography. Hydrophilic interaction liquid chromatography (HILIC) coupled to an electrospray (positive and negative, ES+ and ES−, respectively) single quadrupole mass spectrometer (MS) and a photodiode array (PDA) detector was done on an Acquity UPLC system from Waters (Milford, MA). Computational Methods. The theoretical approach was conducted based on density functional theory (DFT). All geometry optimizations, frequency calculations, and timedependent DFT simulations were performed with the Gaussian 09 program suite. The best-fitting parameters were found to be the ωB97X-D functional and def2-TZVPP basis sets, benchmarked against several functional and basis sets. Ground-state geometries were optimized without constraints at the corresponding level of theory. Frequency calculations were performed to confirm that the geometry was, in fact, a minimum (zero imaginary frequencies) at the corresponding level of theory. TD-DFT calculations34 with the Tamm− Dancoff approximation35 were used to simulate UV−vis spectra of target compounds. MD simulations with a large number of explicit water molecules, in conjunction with TD-

DFT, were used to demonstrate the importance of product− solvent interactions in the accurate assignment of UV−vis transitions. Computational output files are available from the UCSD library digital collections.36 Further experimental details are provided in the Supporting Information (SI).



RESULTS AND DISCUSSION We studied organic byproducts from AuNPs prepared by adding HAuCl4 to pH-adjusted MES solution at the target temperature. Facilitated by N−Au complexation, [AuCl4]− was rapidly reduced by MES, as observed with other amines,20 as evidenced by the color change from the slight yellow color of the precursor to purple/red of the AuNPs (Figure 1a and Video S1). Figure 2 presents the dynamics of the optical properties of the AuNP−byproduct mixture. The peak at 520 nm in Figure 2a is the characteristic of AuNPs and arises from localized surface plasmon resonance (LSPR). Most noticeably, a prominent peak develops at 365 nm, while the characteristic LSPR features of the AuNPs remain unchanged over 5 h. It is thus clear that two distinct phases occur during the synthesis: C

DOI: 10.1021/acs.jpcc.9b03193 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C one is fast and dominated by AuNP-forming reactions and the second is a slower phase during which the AuNP size/shape does not change. Removing the AuNPs from the mixture by filtration at different reaction times (Figure 2b) shows the evolution of several distinct absorption bands and an isosbestic point at 316 nm. Removing the contribution from AuNPs to the spectra in Figure 2a (by subtraction of the initial spectrum, 4 min) reveals additional details of the dynamics and bands that change independently at 200, 235, 300, 360, and 475 nm, Figure 2c,d. We note that Figure 2c agrees well with Figure 2b, supporting the validity of this treatment. When the AuNPs are physically removed by filtration immediately after formation (3 min, Figures 1 and S1), the byproduct absorbance dynamics and composition are changed, Figure 2e,f. The difference in the background below 300 nm is caused by the species (mainly MES) present at 4 min, which are removed by subtraction in Figure 1c. It is clear that the AuNPs affect the product distribution and are involved in the formation of the species absorbing at 360 nm, indicating a catalytic role of the AuNPs. Furthermore, it indicates that the byproducts associate with the AuNP surface since certain byproducts only form when AuNPs are not removed after their formation. Indications of byproduct−AuNP association during the fast AuNP-forming phase are also found in the TEM images, where nonspherical AuNPs (rods/plates) are found when synthesized at 21 °C (Figure 1f), as reported earlier.2 UV−vis spectra of reaction filtrates from a variety of synthesis conditions (pH, reactant concentrations, temperature) were investigated, showing a strong dependence of relative and total concentrations of byproducts, Figure S6. The origin of these absorption peaks has not previously been reported, and we previously observed how AuNP sols derived from the reduction of [AuCl4]− with MES exhibit special optical11 properties that cannot be explained simply by AuNPs or MES-derived N-oxide, as reported for MES oxidation by hydrogen peroxide.37 The current work aims at understanding the reaction and role of organic byproducts. Chemistry of MES Oxidation in AuNP Synthesis. Our experimental results (discussed in the following section) suggest that the reaction involves up to four consecutive oxidation steps of MES. Scheme 1 presents our proposed reaction mechanism, where intermediates confirmed by NMR, MS, and X-ray crystallography are highlighted in blue. Unobserved intermediates in the reaction scheme are believed to be short-lived species necessary for the ultimate formation of oxalic acid. The reaction is initiated by the coordination of [AuCl4]−, or the aqua/hydroxy species such as [AuCl3OH2] and [AuCl3OH]−,38,39 to the nitrogen in MES. The reaction can only be observed at an initial pH above 6 where the nitrogen in MES is deprotonated (pKa = 6.15) and has a free lone pair. In the product mixtures, carbon atoms α and β to the nitrogen are oxidized, and carbon−heteroatom bonds are broken. The first step is the oxidation of MES, either directly to hemiaminal (2) as shown or through an intermediate Noxide as reported previously.40 The iminium ion (4) is formed by dehydration, and loss of a proton forms the dihydro-oxazine (5). Hydration of the enamine leads to two possible isomeric aldehydes (7, 8) that are presumed to be easily oxidized to their corresponding acids (9, 10) under the reaction conditions. Cyclization and further oxidation eventually lead to the N-substituted ethanolamine (13) and free oxalic acid (15). Furthermore, the intermediates presented in Scheme 1

Scheme 1. Proposed Reaction Mechanism of GoldCatalyzed Oxidation of MES to N-(2Sulfoethyl)ethanolamine and Oxalic Acida

a

The structure of purple compounds has been determined by NMR/ HILIC-MS. Blue indicates support from in situ NMR experiments. Oxalic acid has been observed in separate experiments by X-ray diffraction, Figure S23.

may undergo additional oxidation adding to the byproduct distribution. The oxidation steps presented in Scheme 1 can be driven by both Au(III)/Au(I) and dioxygen as oxidizing agents. AuNPs are known to catalyze oxidations with dioxygen as oxidant also after the gold precursor is consumed. In fact, Della Pina et al. demonstrated efficient conversion of tertiary amines to N-oxides under conditions similar to those in our system (2 atm O2,